3 ms·
RE: Rust's pacing I've had a lot of talks with my management about that. For context, I'm on the Cargo team and have authored 11 RFCs (10 approved, 1 pending)
by epage 2y ago
RE: Rust's pacing
I've had a lot of talks with my management about that. For context, I'm on the Cargo team and have authored 11 RFCs (10 approved, 1 pending).
I feel like a lot of the pacing feels slow because:
- As the project matures, polishing whats there takes up a lot of effort
- Conversely, hitting local maximas where things are "just good enough" that individuals and companies don't feel the need to put effort to doing the last leg of work.
- Lack of coordinated teams (formerly Mozilla) doubling down on an idea to hash it out. Hopefully [Project Goals](https://rust-lang.github.io/rfcs/3614-project-goals.html https://rust-lang.github.io/rfcs/3614-project-goals.html) will help a little in this direction.
- As the project has grown, we've specialized a lot more, making it harder to develop a cross-team feature. It takes finesse to recruit someone from another team to help you finish out a cross-team feature. It also doesn't help we've not done a good job developing the cross-team communication channels to make up for this specialization. Again, Project Goals are trying to improve this. In-person conferences starting back up has also been a big help.
As for RFCs, we've been moving in the direction of choosing the level of process thats appropriate for a decision. Unsure how something will look? You just need approval from 2 members of the relevant team to start a nightly only experiment to flesh out the idea in preparation for an RFC. In Cargo, many decisions don't need wide input and are just team votes on an Issue. RFCs drag out when their isn't a someone from the team shepherding it through the process, the RFC covers too much and needs to be shrunk to better focus the conversation, too much is unknown and instead an experiment is needed, or its cross-team and you need to know how to navigate the process to get the vote done (we want to improve this). As for things being approved but not completed, thats a "we need more help" problem usually.
- pdimitar 2y ago> As for things being approved but not completed, thats a "we need more help" problem usually. You know, I would LOVE working on Rust (not just with Rust) and be a part of some of the core team(s). But my impression is that nobody truly has any powerful agency over things and even if you formulate a near-perfect and a PR to go with it, things would still end with several smarter people than me saying "Oh this looks really neat, we should ponder it more and test it further and merge it!" and then it never happens. That, plus I am not sure how is the job stability situation there.